Control Theory in Government’s Economical Decisions

Researchers at K. N. Toosi University of Technology in collaboration with a group of economists from the Iranian National Tax Administration have developed a strategy that could predict government tax revenues for a better budget planning. In a study that outlines this strategy, the researchers found that a model for the tax system needs to be identified, not solely based on similar data in previous years. The traditional models don’t have reasonable accuracy due to the lack of considering other effective factors and uncertainties such as price changes, government policies and the rate of production boom. “If the governments could predict the amount of tax revenue, they could optimally allocate the available resources, so the rate of participation of people in public expenses would be determined – highly important for budget planning!”, says Dr. Hamid Khaloozadeh, professor of Control Systems and the Dean of the Faculty of Electrical Engineering at K. N Toosi University of Technology. It is clear that the tax system of a country is more complicated to be modeled using linear regression-based or classical methods. In addition, such systems are not either linear or stochastic. In fact, they belong to the non-linear systems spectrum which exhibit chaotic behavior. You may think they show totally random behaviors, but despite of their complexity, their behaviors could be modeled.

“Structures such as neural networks are reasonable for modeling such systems, especially when we use Fuzzy logic for modeling the uncertainties. Artificial Neural network Fuzzy Inference System (ANFIS) could be the optimum method we have used for modeling the desired system,” adds Dr. Khaloozadeh.